skills/xsavikx/okf-skills/okf-git/Gen Agent Trust Hub

okf-git

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection attack surface because it processes untrusted data from local Git repositories and commit logs and includes it in its markdown output.
  • Ingestion points: Data is pulled from file contents, file metadata, and Git commit history (authors, dates, and messages) in main.go and relationships.go.
  • Boundary markers: The output markdown files (OKF bundles) do not use distinct delimiters or 'ignore' instructions for the interpolated commit messages or descriptions, which could lead a downstream agent to follow instructions embedded in that data.
  • Capability inventory: The skill can read local Git repositories and write to the local file system (creating markdown files and updating metadata files) via main.go.
  • Sanitization: The skill does not sanitize or escape the content of commit messages or file descriptions before including them in the produced bundles.
  • [EXTERNAL_DOWNLOADS]: The documentation provides instructions to install the connector binary from an external repository.
  • Evidence: The command go install github.com/xSAVIKx/okf-skills/skills/okf-git@v0.1.0 is mentioned in SKILL.md.
  • Context: The repository belongs to the skill author ('xsavikx'), representing a standard installation procedure for this toolset.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 02:02 PM